Impact of stratospheric polar vortex variability on Antarctic surface climate and sea ice

The interannual variability of Antarctic sea ice is considered to be mainly driven by tropospheric and oceanic processes. However, the stratosphere also constitutes a possible source of sea ice variability. The stratospheric variability in the southern high latitudes is dominated by the stratospheric polar vortex (SPV), an extremely cold air mass confined to the pole by strong westerly winds. The SPV is characterized by a large seasonal cycle, peaking in austral winter and breaking down in late spring (with the so-called stratospheric final warming, SFW), but also by interannual variations. While there is robust evidence of a downward impact of the polar stratospheric variability on the Northern Hemisphere surface climate, including sea ice, whether a similar link is present in the Southern Hemisphere is still unsettled. Here, we perform a multi-model assessment of the impact of the dynamical state of the SPV on Antarctic surface climate and sea ice by applying the same experimental protocol to three state-of-the-art general circulation models (GCMs): EC-EARTH, CMCC-ESM and CanESM. The three GCMs have similar ocean and sea ice components but different atmosphere. First, we examine 200-year control experiments and compare them to observations. To assess the impact of the SPV state on the surface and sea ice, we build composites of “strong” and “weak” SPV years based on the late-winter stratospheric conditions. We then compare the anomalous patterns of sea ice concentration during the following spring, as well as anomalies of atmospheric fields such as sea-level pressure and surface temperature. To detect the possible downward stratosphere-troposphere coupling, we also compute the temporal evolution of vertical profiles of zonal-mean zonal wind and temperature. A similar analysis is also carried out using composites based on the timing of the SFW (“early” versus “late”). To further isolate the potential role of the polar stratosphere in driving Antarctic surface climate, we run an additional set of sensitivity experiments with suppressed stratospheric variability. For each model, we build 200-member ensembles of 1-year long runs initialized from the control experiment, with the polar stratosphere nudged to the models' climatology, while the troposphere and the extra-polar stratosphere evolve freely. We then compare the variability of Antarctic sea ice and surface climate in these sensitivity experiments to that of the control run and investigate changes in the suggested mechanisms for the stratospheric downward influence.